27C.A.Estrada,P.K.Nair,M.T.S.Nair,R.A.ZingaroandE.A.Meyers,J.Electroch - PDF document

27C.A.Estrada,P.K.Nair,M.T.S.Nair,R.A.Zi
27C.A.Estrada,P.K.Nair,M.T.S.Nair,R.A.ZingaroandE.A.Meyers,J.Electrochem.Soc.,1994,,802.28M.Lakshimi,K.Bindu,S.Bini,K.P.Vijayakumar,C.SudhaKartha,T.AbeandY.Kashiwaba,ThinSolidFilms,2000,,89.29I.Grozdanov,C.K.BarlingayandS.K.Dey,Integr.Ferroelectr.1995,,205.30M.SavelliandJ.Bougnot,SolarEnergyConversion,inTopicsinAppliedPhysics,ed.B.O.Seraphin,Springer,Berlin,1979,vol.31,p.213.31R.E.Camacho,A.R.Morgan,M.C.Flores,T.A.McLeod,V.S.Kumsomboone,B.J.Mordecai,R.Bhattacharjea,W.Tong,B.K.Wagner,J.D.Flicker,S.P.TuranoandW.J.Ready,,2007,,39.PCCPPaperHowever,ourworkprovidesproof-of-conceptandsuggeststhatthearchitectureandmethodofproductionoerpromise.Wenotethatrecently,theuseofcarbonnanotubecarpetshaveshownpromisewithCdTe/CdSsolarcell;however,theCdTelayerswasontopofthenanotubearraysandtherewasminimalNoattemptwasmadeinthisstudytooptimizethedopingofthesemiconductorlayers.InfutureitwillbeimportanttoensurethattheLPDgrownlayersareuniformincompositionanddoping.Inaddition,itshouldbenotedthataspreparedtheSWNTsamplescontainapproximately15…20%metallictubes,theremainderbeingsemiconducting.Thiswillcontributetothelowerthatdesiredconductivityandhencechargeseparation.Anothertaskyettobecompletedisthedesignofatrans-parenttopcontactthatwillnotincreasemanufacturingcosts.Conductivematerialisrequired,inordertocollectchargecarriersacrosstheentiretyofthetopsurface.Thetopcontactmustalsobeopticallytransparentsothatlightreachestheabsorberlayer.Materialssuitableforourdeviceincludepoly(3,4-ethylenedioxythiophene)…poly(styrenesulfonate)(PED-OT:PSS)andtransparentconductiveoxides(TCOs)suchasITOorfluoride-dopedtinoxide.OtherpotentialmaterialsareSWNTorgraphenepolymercomposites.AcknowledgementsThisworkwassupportedbytheRobertA.WelchFoundationNotesandreferences1H.J.Moller,SemiconductorsforSolarCells,ArtechHouse,Boston,1993.2M.Pagliaro,G.PalmisanoandR.Ciriminna,FlexibleSolar,Wiley-VCH,Weinheim,2008.3P.J.F.Harris,CarbonNanotubesandRelatedStructuresCambridgeUniversityPress,Cambridge,1999.4B.J.Landi,R.P.Raaelle,S.L.CastroandS.G.Bailey,Photovolt.:Res.Appl.,2005,,165.5S.Cataldo,P.Salice,E.MennaandB.Pignataro,Environ.Sci.,2012,,5919.6E.Kymakis,M.M.Stylianakis,G.D.Spyropoulos,E.Stratakis,E.KoudoumasandC.Fotakis,Sol.EnergyMater.Sol.Cells,2012,,298.7S.De,P.E.Lyons,S.Sorel,E.M.Doherty,P.J.King,W.J.Blau,P.N.Nirmalraj,J.J.Boland,V.Scardaci,J.JoimelandJ.N.Coleman,ACSNano,2009,,714;G.Fanchini,S.Miller,B.B.ParekhandM.Chhowalla,,2008,,2176;A.A.GreenandM.C.Hersam,,2008,,1417.8V.SgobbaandD.M.Guldi,J.Mater.Chem.,2008,,153;D.M.Guldi,G.M.A.Rahman,F.ZerbettoandM.Prato,Chem.Res.,2005,,871;D.M.Guldi,G.M.A.Rahman,N.Jux,N.TagmatarchisandM.Prato,Angew.Chem.,Int.,2004,,5526;S.Roy,R.Bajpai,A.K.Jena,P.Kumar,N.KulshresthaandD.S.Misra,EnergyEnviron.Sci.,2012,,7001;N.J.Alley,K.-S.Liao,E.Andreoli,S.Dias,E.P.Dillon,A.W.Orbaek,A.R.Barron,H.J.ByrneandS.A.Curran,Synth.Met.,2012,,95.9M.W.Rowell,M.A.Topinka,M.D.McGehee,H.J.Prall,G.Dennler,N.S.Sariciftci,L.B.HuandG.Gruner,Phys.Lett.,2006,,233506.10E.T.Mickelson,C.B.Human,A.G.Rinxler,R.E.Smalley,R.H.HaugeandJ.L.Margrave,Chem.Phys.Lett.,1998,296,188.11J.L.Delgado,M.A.HerranzandN.MartšJ.Mater.Chem.,1417.12A.Mazzoldi,D.DeRossiandR.H.Baughman,Proc.SPIE,25;U.Vohrer,I.Kolaric,M.H.Haque,S.RothandU.Detla-Weglikowska,,2004,,1159;L.J.Sweetman,L.Nghiem,I.Chironi,G.Triani,M.inhetPanhuisandS.F.Ralph,J.Mater.Chem.,2012,,13800.13Z.Niu,W.Zhou,J.Chen,G.Feng,H.Li,W.Ma,J.Li,H.Dong,Y.Ren,D.ZhaoandS.Xie,EnergyEnviron.Sci.,1440.14P.J.SebastianandP.K.Nair,Adv.Mater.Opt.El

ectron.,211;P.J.Sebastian,A.SanchezandP.
ectron.,211;P.J.Sebastian,A.SanchezandP.K.Nair,Mater.Opt.Electron.,1993,,133;P.J.SebastianandH.Hu,Adv.Mater.Opt.Electron.,1994,,407;P.J.Sebastian,H.HuandA.M.FernaAdv.Mater.Opt.Electron.,1995,,11.15T.Rattanavoravipa,T.SagawaandS.Yoshikawa,Sol.EnergyMater.Sol.Cells,2008,,1445;C.-J.Huang,M.-P.Houng,Y.-H.WangandW.-J.Chang,Jpn.J.Appl.Phys.,1998,,L158;L.P.Deshmukh,P.P.HankareandV.S.Sawant,Sol.Cells,1991,,549.16G.Hodes,ChemicalSolutionDepositionofSemiconductor,MarcelDekker,NewYork,2003.17B.J.Landi,C.M.Evans,J.J.Worman,S.L.Castro,S.G.BaileyandR.P.Raaelle,Mater.Lett.,2006,,3502;B.J.Landi,S.L.Castro,H.J.Ruf,C.M.Evans,S.G.BaileyandR.P.Raaelle,Sol.EnergyMater.Sol.Cells,733.18R.LoscutovaandA.R.Barron,J.Mater.Chem.,2005,,4346.19J.Nelson,ThePhysicsofSolarCells,ImperialCollegePress,London,2003.20I.W.Chiang,B.E.Brinson,A.Y.Huang,P.A.Willis,M.J.Bronikowski,J.L.Margrave,R.E.SmalleyandR.H.Hauge,J.Phys.Chem.B,2001,,8297.21V.M.Garcia,M.T.S.Nair,P.K.NairandR.A.Zingaro,Semicond.Sci.Technol.,1996,,427.22D.J.FloodandA.R.Barron,ASimpleTestApparatustoVerifythePhotoresponseofExperimentalPhotovoltaicMaterialsandPrototypeSolarCells,ConnexionsWebsite.http://cnx.org/content/m42271/1.3/,Nov.18,2012.23V.C.Moore,M.S.Strano,E.H.Haroz,R.H.Hauge,R.E.Smalley,J.SchmidtandY.Talmon,NanoLett.,2003,,1379.24M.T.S.Nair,P.K.Nair,H.M.K.K.Pathirana,R.A.ZingaroandE.A.Meyers,J.Electrochem.Soc.,1993,,2987.25A.Kylner,J.LindgrenandL.Stolt,J.Electrochem.Soc.,1996,,2662.26V.M.Garcia,P.K.NairandM.T.S.Nair,J.Cryst.Growth,113.PaperPCCPThepresenceofcadmiumatthesamplesurfaceismostlikelyduetocadmiumspecies(hydroxides,carbonates)previouslyadsorbedonthesurfaceoftheCdSefilmarere-dissolvedintheCuSeLPDbathandthensubsequentlyre-depositedontotheCuSewindowlayer.Anotherexplanationisthatmetalsaresignificantlymobilewithinthefilms.Infact,thishasbeenshownpreviously:CdSefilmshavebeendopedwithcopperbysimpleimmersioninaqueousCusolutions.Additionally,inCdSfilmscadmiumcanbeexchangedwithcopperquantitativelydowntoadepthof300nmbyanaqueousbathmethod.However,theseexplanationswouldbeexpectedtoresultinsmallamountsofCdthroughoutthefilmorclosetotheCuSe/CdSeinterface,notatthetopsurface.Whateverthesourceofthecadmium,itisnotexpectedtohaveasignificanteectonthefunctionofPVdevices.XRDwasperformedinordertodeterminethephaseofCuSepresent.Thefilmhassomecrystallinity,asweakdiractionwasseenwithpeaksmatchingbulkberzelianiteCuSe,furthercorroboratingXPSdata(Fig.10).Asafinalcharacterization,depth-profileRamanspectrawerecollectedonthefilms.CuSeandCdSehavetheirstrongestRamanmodesat263and209cmrespectively.Itmustbementionedthatthesefrequenciesarecoincidentwiththeradialbreathingmodes(RBMs)ofSWNTsfoundbetween150and250cm.Thus,caremustbetakenwhenassigningpeaks.Additionally,Ramandepthprofiledatasuersfromlimitations.Scatteringoccursfromaboveandbelowthefocalplane,thereforedepthslicesoverlap.Also,deeperslicessuerfromreducedsignaltonoiseduetosignificantscatteringfromthesampleabovethefocalplane.Ramanspectrawerecollectedwiththemicroscopefocalplanevariedfrom2mabovethesamplesurfaceto2belowin1msteps.TheresultingspectrashownoneoftheexpectedpeaksattributabletoSWNTs(D,Gmodes)andcon-tainasingle(CuSe)peakat263cm.Spectracollectedwiththefocalplaneatthesurfaceand1mbelowshowaveryslightpeakbegintoappearat209cm,indicatingCdSe(Fig.11).Unfortunately,at2mdepthandbelow,baselinenoiseincreasedsothatnopeakswereobserved.ScatteringexpectedforSWNTswasnotseeninanyspectrum,indicatingthatthesemiconductorcoatingsaremicronsthick.3.5.PVtestingDeviceswe

resubjectedtotestingforphotovoltaicrespo
resubjectedtotestingforphotovoltaicresponse.Itshouldbenotedthatthedevicesdescribedherehavenotyetbeenoptimizedintermsofrelativeorabsolutefilmthicknessofthemetalselenides.Additionally,as-prepareddeviceslackatransparenttopcontact.Fortestingpurposes,silveradhesivepaintwasadequatetoprovideproofofconcept(Fig.2g).Fig.S5(ESI)showsthetestingconfiguration.Thegoalofthistestwastodetermineiftheconceptisviableratherthancreateaworkingdevice.Devicetestingwasperformed(topcontact:silverpaintspots)whichconfirmedPVfunctionality,albeitwithminimaleciency.ThecharacteristicIVcurveforthetestdeviceisshownbelow(Fig.12).Theopencircuitvoltage()wasfoundtobe1.28mV;shortcircuitcurrent()was4.85A.Foradevicetobeviable,willneedtobeincreasedbynearly3ordersofmagnitude;likewiseshort-circuitcurrentshouldbeinthemArange.Nevertheless,webelievethatgiventheun-optimisednatureofthedevicewithregardtofilmthickness,topandbottomcontacts,andthefilmdopinglevelsasdefinedbythecomposition,theabilitytocreateaactivedevicethroughsimpleLPDprocessingoershopethatinorganicthinfilmdevicescanbepreparedthroughsimplechemical4.ConclusionsWehavedemonstratedafunctional,albeithighlyinecient,thinfilmsolarcellproducedbyaLPDmethod,whichcanbemanufacturedonflexiblesubstrates.Manyimprovementsarestillneededintermsofdesign,performance,andproduction.Inparticular,furthercharacterizationofsemiconductorfilmsisneededtofindoptimalthicknessandtoimprovecrystallinity.Fig.10XRDpatternofCuSe…CdSeBuckypaper.Thesimulatedpatternforberzelianite(JCPDSCardNo.06-0680)isshown.Fig.11Depth-profileRamanspectra(514nmexcitation)ofCuSe…CdSeBucky-paper,from+2tomrelativetosamplesurface.Fig.12curvefortestdevice(=0.985).PCCPPapercompoundingtheformationofcracks.Inordertolearnwhethersamplehandlingwascausingfilmcracking,asmallpieceofBuckypaper-on-tapewasaxedtoanaluminiumSEMsamplestub.ThestubwascoatedwithrubbercementtoprotectitfromtheLPDsolution,andthenCdSewasdepositedontheimmobilizedBuckypaper.SEMrevealed,again,alargeamountofcrackingoftheCdSefilm.Therefore,crackingisnot(atleastnotentirely)causedbysamplesflexingduringhandling.Itisstillunclearwhethertheproblemisduetointernalstresseswithinthefilmorsomeotherfactor.ToovercomethisproblemshorterdepositionswerestudiedonBuckypaper-on-tapesamples.Severecrackingoccursafter12hofdeposition.At4and8hours,uniformfilmswithsmoothtopsurfacesareformed(Fig.8andFig.S3,ESI).WhilesomecracksarestillvisibleinSEM,theyaremuchfewerandsmaller.XRDwasattemptedonCdSeSWNTcomposites;howeverthesamplesdidnotdiract.Eitherthematerialistrulyamorphous,orthecrystallinedomainsaresmallerthanca.4nm.However,previousXRDstudiesofsimilarLPDCdSefilmsdepositedonglassrequiredannealingbeforeanyweakdiractionwasobserved.3.4.DepositionofCuSewindowlayerCuSewindowlayerwasdepositedontopoftheCdSelayer(Fig.2f).However,initialattemptstodepositcopperselenideusingDMSUandcitrateat60Cwerenotreproducible;thefilmswereinconsistentinboththicknessandquality.AllpreviousCuSeLPDmethodsusingDMSUhaveresultedinfilmsoftheklockmannitephase(CuSe).ThedesiredphaseforPVwindowlayersisberzelianite(CuSe,where0.2),asithasbetterconductivity.Therefore,analternateseleniumsourcewasexamined,sodiumselenosulfate(Na).Naconvenientlypreparedbyheatingelementalseleniuminanaqueoussolutionofsodiumsulphite.Itisunstable(thoughlesssothanDMSU)andmustbekeptfromoxygentoavoidbulkprecipitationofelementalSe.SeveralselenosulphateCuSepreparationswereattempted,usingcitrateorNTAascoppercomplexants.AlsotestedwerereactionschemesemployingonlyNHasligand.methodofGarciaetal.p

rovidedthebestqualityfilms.SEMimagesofth
rovidedthebestqualityfilms.SEMimagesoftheCuSefilmsshowcolloidalmaterialandsmallcrystallitesonthesurface(Fig.8).DeterminationofCuSefilmcompositionandthicknessonCdSeBuckypaperswasnontrivial.Equivalentfilms(16hdeposition)grownonglasshaveathicknessof200nm.datawascollectedtofindaveragefilmstoichiometry.Thepresenceofasmallamountofcadmium(3%)wasdetected.EDSsamplingdepthisontheorderof1m;thepresenceoftracecadmiumimpliesthatCuSefilmsareinfactapproxi-mately1mthick.AnalysisbyXPSalsoshowedthepresenceofcadmium.2%).SinceXPSisasurfacetechnique,thiscloseagreementbetweenXPSandEDSwasconfusing.ItwasassumedthatCdinEDSresultedfromtheunderlyingabsorberlayer;cadmiumwasnotexpectedtobepresentatthesurface.Takentogether,XPSandEDSsuggestedthateitherCuSefilmsaredopedwithsmallamountsofcadmium,orthatCuSecoverageisincomplete.SEMimagestakenoverlargeareasshowedunifor-mityofthesurfacetexture;thusthefilmsappeartobehomo-geneous(Fig.9).FurtherinformationwasobtainedbyXPSdepthprofiling.Iterativeionsputteringandelementalanalysisshowedcadmiumpresentonlyatthesurfaceofthefilm.Uponsputteringintothesamplecadmiumcontentdisappearscompletely(asdoesoxygen),leavingaconsistentstoichiometryofCuSethroughoutthefilm.ItwasnotpossibletosputterthroughtheCuSewindowlayerdownintotheCdSeabsorber.Finally,baseduponXPSandEDS,itappearstheCuSewindowlayerisofthedesiredberzelianitestoichiometry,hasathicknessofm,andiscontaminatedatthesurfacebycadmium.Fig.7SEMimageofCdSe(16hdeposition)onBuckypaper-on-tape.Fig.8SEMimageofCdSefilm(4hdeposition)onBuckypaper-on-tape.Fig.9SEMimageofCuSefilmonCdSeonBuckypaper-on-tape.PaperPCCP1.5:10hadnoeectonfilmcomposition.Conversely,ashydrolysisofDMSUisbelievedtobetherate-limitingstepforLPDofCdSefilms,thesulphite:Seratiowasalsoloweredto1:20.Itwasthoughtthatthiswouldincreasethedepositionrate,resultinginbetterCdSefilms.Unfortunatelythisapproachfailed;depositedfilmswereevenhigherinselenium.Finally,theratioofcitratetocadmiumwaslowered;therationalebeingthatincreasedconcentrationsofCd(OH)(assuminghydroxideclustermechanism)orfreeCdion)shouldconsumeSebeforeitcanbeoxidizedtoSeRatioswerevariedbetween2:1and4:1.Forcitrate:Cdratiosbelow3:1,visibleCd(OH)formedimmediatelyonadditionofammonia,andbelow2.5:1filmdepositionwaspoor.Changingthisparameterdideectivelyalterfilmstoichiometry(Table1).Acitratetocadmiumratioof3:1producesfilmscontainingminimalexcessselenium.ThephotoconductivityofTheCdSefilmwastestedpriortocellfabrication(Fig.S1,ESI).Thisallowsfortherapidscreeningofmaterialsorsynthesisvariableofasinglematerialevenbeforeissuesofcelldesignandconstructionareconsidered.Thevoltagewasmeasuredwithashutterheldoverthesampleandwiththeshutterremoved.Thedierenceinvoltageisadirectindicationofthechangeinthephoto-conductanceofthesample.AsmaybeseenfromFig.5thereisadistinctphoto-responsefortheCdSe.3.2.FabricationofbackcontactsAttemptstoperformelectricaltestsontheCdSefilmsonBuckypaper/PTFEwerehamperedbydicultygainingelectricalcontacttothebacksurfaceoftheSWNTs.Severalremedieswereattempted,includingsputteringofgoldontothebackoftheBuckypaperafterpeelingitfromtheTeflonmembraneorusingconductivesilveradhesive.VerythinBuckypapersubstratesaretoofragiletoberemovedfromthebackingmembranes,soadierentsolutionwasneeded.BypaintingthetopsurfaceoftheBuckypaperwithconductivesilveradhesivewhileitisundersuctioninafiltrationapparatus,itispossibletopartiallyimpregnatesilverwithintheSWNTnetwork(Fig.2aandb).Caremustbetaken,however,nottousetoomuchsilverortoleavethevacuumontoolong.Ifsilverbleedsthrou

ghtotheundersideoftheSWNTs,cleanseparati
ghtotheundersideoftheSWNTs,cleanseparationoftheBuckypaperfromtheTeflonmembraneisnotpossible.AfterthesilverpaintdriedtheBuckypaperwasinvertedandcarefullytransferredtotheadhesivesideofastripofPETpackingtape(Fig.2c).Thetapehadbeenperforatedinasmallarea,correspondingtothepositionofthesilveronSWNTs.Nextthefiltermembranewascarefullypeeledaway,transfer-ringtheSWNTstotheadhesivetape(Fig.2dandFig.S2,ESI).BeforesubjectingtheBuckypapertoLPD,holesinthepackingtapeweremaskedbyrubbercementtopreventCdSefromdepositingonthebackcontacts.ResultingBuckypapershaveaverysmoothtopsurfaceasseenbySEM(Fig.6).Inadditiontheyareflexible,robustanddonotdelaminatefromthetape,evenundertheconditionsofLPD.3.3.Second-generationCdSeBuckypaperDepositionofCdSeonBuckypaper-on-tape(Fig.2e),usingtheconditionsoptimizedabove,yieldedsurprisingresults.After16hoursofdeposition(44h),SEMimagingrevealedthatthefilmswereseverelycracked(Fig.7).SEMfurthershowedthatthefilmsweremuchthickerthanthosefrompreviousBuckypapersamples.OneexplanationforthisobservationcouldbethatBuckypaper-on-packingtapesampleshaveasmoothertopsurfacethatallowsrapidnucleationandthere-forefasterfilmgrowth.Alternately,itispossiblethatSWNTsontapeareofhigherdensityatthedepositionsurfacesincetheyhavebeeninverted(andalsocompressedwithsucientforcetoadherethemtothetape),andthereforelesstimeormaterialisrequiredtoinfillthevoidspace.Theobservedextensivecrackingisproblematic,asanydirectcontactofSWNTstothewindowlayerwillshorttheheterojunction.Crackingcouldbeattributedtoexcessiveflexingofthesubstrate,eitherduringLPDorsubsequentSEMsamplepreparation.Thickfilmsarelikelytobemorebrittle,Table1Eectofcitrate:cadmiumratioondepositedCdSefilmstoichiometry(asdeterminedbyEDS)Citrate:cadmiumratioCd:Sefilmstoichiometry4.0:11:1.73.5:11:1.43.0:11:1.22.5:11:1.32.0:1NofilmFig.5PhotoresponseofaBuckypapercoatedwithLPDgrownCdSe.Fig.6SEMimageofBuckypaper-on-tape(SWNTside).PCCPPaperconcentrationsofCdandSearecontrolledbythereactionsshownineqn(4)…(6).+SeCdSe(1)+2OH+SeCdSe+2OHL(4)NC(NH)Se+OHHSe+MeNCN+HO(5)+OHO(6)Itisdiculttodeterminewhichmechanismisdominantforagivenfilmdeposition.Infact,bothareoperativetosomedegreeundermostreactionconditions.Therearesomeconditionsthatfavouronemechanismtheother.Ingeneral,thehydroxideclustermechanismcanbefavouredbyincreasingpHorbydecreasingtheratioofligandtocadmium.IncreasingpHacceleratestheformationofhydroxideclusters(byincreasedOHconcentration);reducingtheamountofcomplexantdoesthesamebyincreasingavailablefreeCdUsuallytheion-by-ionmechanismgivesslowerdepositions,butproducedfilmshavelargerterminalthicknesses.whetherduetodepositionrateorotherfactors,ion-by-ionfilmscommonlyhavelargercrystallites.3.1.DepositionofCdSeabsorberlayerCdSefilmsweredepositedonBuckypaperfromCdSOsodiumsulphite-stabilizedDMSU,usingsodiumcitrateastheCdcomplexingligand,inaqueousammoniasolution(pH10.4)atroomtemperature.Undertheseconditionsitisexpectedthatthehydroxideclustermechanismisdominant.FilmsdepositedontoSDS-derivedBuckypapersonPTFEmembranes(12hdepositiontime)wereshownbySEMtohaveincompleteinfillofthevoidspacebetweenSWNTs.Inaddition,largenumbersofcolloidalCdSeparticleswerepresentonthefilmssurfaces(Fig.4a).SincetheproductionofanecientPVdevicedependscriticallyonthequalityofthep…nheterojunction,aplanarinterfaceisdesired,freefromdefectsordebris.Further-more,ifcracksintheCdSelayerexposeSWNTtosubsequentdepositionoftheCuSelayer,andhencewillresultinashortcircuitofthedevice.Reductionofthereagentconcentrationshadnoeectonimprovi

ngthequalityofthefilms;infactwhenthebath
ngthequalityofthefilms;infactwhenthebathconcentrationisreducedthedepositionrateonthesubstrateisretardedwithnoconcomitantslowingofbulkprecipitation(i.e.,thinnerfilms,samecolloidaldebris).Performingmultiple4hdepositionsfromfreshbathsproduceduniformfilmswithcleansurfacesalthoughcompleteinfillrequired16totalhours(Fig.4b).X-rayphotoelectronspectroscopy(XPS)analysisshowedthestoichiometryofthefilmsproducedtocontainanexcessofcadmium,havingCd:Seratiosofbetween1.7:1and1.9:1.Energy-dispersiveX-rayspectroscopy(EDS)confirmedtheXPSdata.Furtheranalysisshowedalargeamountofoxygenpresent.Itisunderstoodthatoxygenchemisorbstometalchalcogenidefilms;moreover,knownimpuritiesinLPDfilmsincludecadmiumhydroxidesandcadmiumcarbonates.Thepresenceoftheseimpuritiesexplainstheimbalancebetweencadmiumandselenium.Degassedwaterwassubsequentlyusedtoprepareallsolutionsandresultedinfilmsofnegligibleoxygencontent.ThestoichiometryofCdSefilmsproducedwithdegassedwaterwasfoundbyEDStobeseleniumrich;Cd:Seratioswere1:1.7.Thedisparityisattributedtotheformationofzerovalentselenium,whichprecipitatesfromsolutionasablackorredsolidandlikelybecomesentrappedinthefilm.Lowerconcentrationsofseleniumrelativetocadmiumonlyresultedinsloweddepositionandhadnoeectonfilmcomposition.Theratioofstabilizertoseleniumwasvariedasasecondexperiment.SodiumsulphiteisareducingagentusedtoslowtheoxidationofDMSUandpreventtheformationofbulkselenium.Increasingtheratioofsulphite:Sefrom1:10toFig.3SEMofBuckypaperpreparedbyfiltrationofSDSSWNTdispersion.Fig.4SEMimagesofCdSefilmsonBuckypaper:depositiontimeswere12h(a)and44h(b).PaperPCCPsubjectedtoprobesonicationfor60minutes.TheresultingSWNTdispersion(40mL)wasvacuumfilteredoveraPTFEmembrane(47mmdiameter,0.2mporesize),,Fig.2a.Afterallsolventhadbeenremoved,conductivesilverpaintwasbrushedontoanareaatthecentreoftheBuckypaper(Fig.2b).Thevacuumwasleftonfor2minutesinordertoallowsilvertopartiallyinfiltratetheSWNTs.Thepaperwastheninvertedandpressedontoapieceofpolyesterpackingtape(3MPET),whichhadbeenperforatedwithseveralholesbyaneedleinasmallarea(Fig.2c).CarewastakentoensurethatthesilverpaintwasmatchedtotheperforatedareaofthetapetoallowelectricalcontacttobemadefromthebackoftheassembledBuckypaper-on-tape.Beforefilmdeposition,theperforationsinthetapewerecoatedwithathinlayerofrubbercement(Elmers)toprotectthebackcontactsfrombathsolutions.2.2.CdSedepositionToaglassjarwasaddedCdSO(6.0mL,0.10M),sodiumcitrate(3.0mL,0.60M),NHOH(2.4mL,1.50M),deionizedwater(3.4mL),andfreshlypreparedstabilizedDMSUsolution(5.2mL,0.10M).Moistsubstrates(immersedinwater5minutesbeforedripdrying)wereimmersedverticallyinordertominimizebulkprecipitateontheirsurfaces.Thejarwascappedandthereactionwasallowedtoproceedfor4hoursatroomtempera-ture.Whenthedepositionwascomplete,sampleswereremovedfromthebathandrinsedwithdeionizedwater,thenplacedinavacuumdesiccatortodry.2.3.CuSedepositionToaglassjarwasaddedCuSO(5.0mL,0.20M),NH(3.0mL,1.50M),deionizedwater(7.0mL),andNa(5.0mL,0.18M)Moistsubstrates(immersedinwater5minutesbeforedripdrying)wereimmersed,verticallyinordertominimizebulkprecipitateonthesurfaces.Thejarwascappedandthereactionwasallowedtoproceedfor16hoursatroomtemperature.Whenthedepositionwascomplete,sampleswereremovedfromthebathandrinsedwithdeionizedwater,thenplacedinavacuumdesiccatortodry.3.ResultsanddiscussionBuckypapersubstrateswerepreparedbysimplevacuumfiltrationofSWNTdispersionsthroughPTFEmembranes.Buckypaperspreparedfrom-dichlorobenzene(ODCB)SWNTdispersionswerefoundtohavepoorconductivity(.4

00Scm).Filmsofadequateconductivityareobt
00Scm).Filmsofadequateconductivityareobtainedonlyafterhightemperaturevacuumannealing.SeveralaqueousdispersionsofSWNTsusingvarioussurfactantswerethenexamined.Buckypapersproducedfromsodiumdodecylsulphate(SDS),sodiumdodecylbenzene-sulphonate(SDBS),anddodecyltrimethylammoniumbromide(DTAB)showedconductivitiesof1500,500,and800Scmrespectively.Itshouldbenotedthatconductivityisdependentonfilmthickness,however,SEMshowedallsamplestohavethicknessesof2…3m.SDSisknowntoeectivelysuspendcarbonandisthemostcommonlyusedsurfactantforSWNTs.WehaveattributedtheimprovedconductivityofBuckypaperscastfromSDStobeasaconsequenceofamorehomogeneousstartingdispersion.Fig.3showsatypicalBucky-paperdepositedfromSDSsolution.AgeneralschemeforLPDofCdSefilmsinvolvesthehydro-lysisof-dimethylselenourea(DMSU)inalkalineconditionsinthepresenceofacadmiumsalt.AcomplexingligandisalsorequiredinordertopreventprecipitationofCd(OH)aswellastoregulatetheconcentrationoffreeCdionsinsolution.Typicalcomplexingagentsaresodiumcitrateortartrate,triethanolamine,ethylenediamine,ornitrilotriaceticacid.AlthoughonitsfaceLPDappearstobeaverysimpleprocess,infactitisquitecomplexasanumberofequilibriumandreactionsareinvolved.Smallchangesinreactionconditionscanhavedramaticeectsonthequalityandpropertiesofproducedfilms.LPDofmostII…VIsemiconductorfilms,includingCdSe,isconsideredtooccurbyeitheroftwopossiblemechanisms:ion-by-ionorhydroxidecluster.Eqn(1)givestheion-by-ionmechanism.Similarly,thehydroxideclustermechanismisdescribedbyeqn(2)and(3).Forbothmechanisms,theFig.2Schematicrepresentationofthefabricationstepsemployed:(a)deposi-tionofSWNTsontoaTeflonmembrane,(b)paintingsilverpastecontacts,(c)attachmentofadhesivetapetomaskonesideoftheBuckypaperwithalignmenttothesilvercontacts,(d)removaloftheTeflonmembrane,(e)LPDgrowthofCdSe,(f)LPDgrowthofCuSe,and(g)paintingsilverpastecontacts.PCCPPaperhighsurfacearea(1600mandtheirelectron-acceptingproperties,meanthatSWNTshavepotentialasbothaconductivenetworkingeneral,andaspromotersofexcitiondissociation.Infact,forthesereasonscarbonnanotubesingeneralhavebeeninvestigatedasabaseextractionlayerorholetransportlayersorganicsolarcells.ThereforeSWNTsareanidealcandidateforuseinPVapplications;infacttheyarecurrentlybeingstudiedforuseintransparentelectrodesaswellasintheactivelayeroforganicphotovoltaic(OPV)devices.SWNTsaremechanicallyflexible,allowingthemtobedepositedontoflexiblesubstratesbyfiltration,spraying,orprinting,providinganelectricallyconductivematrixontowhichsemiconductorfilmscanthenbedeposited,creatingaembeddedconductionpathwaywithinanabsorberlayer(Fig.1c).TheeasiestmethodforcreatingaSWNTsuitableforgrowthofathinfilmisintheformofso-calledBuckypaper.10,11CommonlypreparedbyfiltrationofSWNTsuspensions,ithasalreadybeenusedinfuelcellandcapacitorapplications.ApromisingroutetoinexpensivePVdevicesisliquidphasedeposition(LPD)alsoknownaschemicalbathdeposition(CBD).Thissolutionphasemethodallowsthedepositionofavarietyofsemiconductorthinfilmsunderambientconditionsusingsimpleequipmentfoundinanywetchemistrylaboratory.LPDiswellsuitedtothepreparationofconformalfilmsonflexiblesubstrates,suchasBuckypaper.Raaelleandco-workershaveshownthatCdSequantumdotscanbeattachedtoSWNTsandusedinpolymersolarcellsasaroutetoexcitondissociationandcarriertransport.Bycontrast,previousresearchinourgroupfocusedondepositionof(12…16)semiconductormaterials(CdSandCdSe)ontoSWNTsinsurfactantsolutions.ThisstudyshowedthatuniformcadmiumchalcogenidefilmscouldbepreparedonSWNTs.HereinwedemonstrateLPDprepa

rationofcadmiumandcopperselenidefilmsonp
rationofcadmiumandcopperselenidefilmsonpreformedSWNTfilms,oftencalledBuckypapers.CadmiumselenidefilmsdepositedbyLPDaren-typewithadirectbandgapof1.74eV,whilecopperselenideLPDfilmsarep-typewithabandgapof2.26eV.SWNTPVdevicesdescribedhere,usingcadmiumselenideastheabsorbermaterialandcopperselenideasthetop(window)layer,couldhave,theoretically,ecienciesapproaching20%.ThearchitectureofourproposeddevicesconsistsofBuckypaperatthebackofthecell,coveredandinfiltratedwithCdSe,whichextendsinacontinuouslayerabovetheSWNTs.CoatingtheCdSewithalayerofCuSeformsabulk-heterojunction(Fig.1c).Thegoalofthispresentstudyistodeterminewhethersuchadevicearchitectureoerpotential.Theresultsoftheseinitialexperimentsarepresentedherein.2.ExperimentalDeionizedwaterwasdegassedpriortousebybubblingwithargonfor15minutes.CdSO,CuSOO,NHOH(28…30%),,sodiumcitrate,sodiumdodecylsulphate(SDS,(SDS,12H25SO4]),andelementalseleniumwerepurchasedfromSigmaAldrichandwereusedasreceived.-Dimethylselenourea[DMSU,Se]waspurchasedfrom3BScientificCorp.Single-walledcarbonnanotubes(HiPcoprocess)wereobtainedfromtheCarbonNanotubeLaboratoryatRiceUniversity,andwerepurifiedfollowingtheliteratureprocedurepriortouse.StabilizedDMSUsolutionwaspreparedbydissolvingsodiumsulfite(13mg,0.10mmol)indeoxygenated,deionizedwater(10mL).-Dimethylselenourea(151mg,1.0mmol)wasthenaddedtothesulphitesolutionwithstirringunderargon.DMSUsolutionsdecomposewithin1…2daysevenifstoredunderargon.Sodiumselenosulphatesolutionwaspreparedbydissolvingsodiumsulphite(5.04g,0.04mol)indeionized,deoxygenatedwater(100mL).Seleniumpowder(1.58g,0.02mol)wasthenadded,andthemixturewasheatedto60Cwithstirringfor2hunderargon.After2hours,residualSewasremovedbyfiltration.Selenosulphateconcentrationisestimatedatca.0.18M.Solutionsareusablefor3…4daysifstoredunderargon.XPSdatawereacquiredonaPhysicalElectronics,Inc.PhiQuanterainstrumentwithmonochromatedAl-KX-raysource.SEMimagingwasperformedwithaFEIQuanta400high-resolutionfieldemissionscanningelectronmicroscope(acceleratingvoltageof30kV)equippedwithliquidN-cooledEDSdetector.XRDdatawereacquiredusingaRigakuD/MaxUltimaIIconfiguredwithaverticaltheta/thetagoniometer,Cu-Kradiation,andgraphitemonochromator.RamandatawereobtainedusingaRenishawRamanmicroscopewitha514nmlaserexcitationsource.Conductivitydatawerefoundusingafour-pointprobeatacurrentof0.5mA.ThephotoconductanceoftheCdSe/Bucky-paperandthePhotoresponseoftheCuSe/CdSe/Buckypapersamplesweretestedusingliteratureprotocols.AschematicrepresentationofthedevicefabricationstepsisshowninFig.2.ItshouldbenotedthatitisimportantfortheSWNTstoonlybeimbeddedintothelowerCdSeabsorberlayer.IfSWNTsremainuncoatedpriortodepositionoftheCuSewindowlayerthecellwillbeshort-circuited.2.1.BuckypapersubstratesBuckypaperwaspreparedbyavariationofliteraturemethods.10,12SWNTs(10mg)weredispersedin1%SDSsolution(100mL)andFig.1Schematicdiagramof(a)traditionalthinfilm(CdTe)solarcell,(b)athinfilmsolarcellinwhichthebackcontactisembeddedintotheabsorberlayer,and(c)apotentialSWNT-basedthinfilmPVdevice.PaperPCCPThinfilmCdSe/CuSephotovoltaiconaflexiblesinglewalledcarbonnanotubesubstrateChristopherE.Hamilton,DennisJ.FloodandAndrewR.Barron*Liquidphasedeposition(LPD),usingCdSO-dimethylselenourea,hasbeenusedtogrowCdSeabsorberlayerontosinglewalledcarbonnanotube(SWNT)derivedbackcontactsubstrates.Thenanotubesareimbeddedin,andpenetrateinto,theCdSeabsorberlayerforthegoalofenhancingexcitiondissociationandcarriertransport.TheCd:Sefilmstoichiometryvariedbetween1:1.7to1:1.3dependingonthedeposit

ionconditions.TheCdSe/SWNTlayersshowappr
ionconditions.TheCdSe/SWNTlayersshowappropriatephotoresponse.LPDwasalsousedtogrowaCuSewindowlayerontowhichsilvercontactsweredeposited.TheresultingPVdeviceshowsacharacteristiccurve.Despiteboththeopencircuitvoltage=1.28mV)andshortcircuitcurrent(=4.85A)beinglow,theresultingdeviceissuggestiveofthepossibilityoffabricatingaflexiblethinfilm(inorganic)solarcellbysolutionprocesses.1.IntroductionSolargenerationofelectricityhasbeenanactiveareaofresearchsinceshortlyafterBecquerelfirstdiscoveredthephotoelectriceectin1839.Concernsoverfossilfuelproductionaswellasissuesofpollutionandclimatechangehavespurredintensiveresearchintophotovoltaics(PV)aswellasotherrenewableenergysources.Solaristhemostutilizedofthegreenenergysourcestodate.However,ithasstalledwithregardtomarketpenetration.GiventhatthetotalsolarenergyadsorbedbytheEarthisca.Jperday,whichisenoughtomeettheworldsenergyneedsforayear,sunlightisnotalimitationofadoption.Atpresentthemainbarriertomassadoptionisthecostincomparisonwithhydrocarbonandcoal,inparticulartheincreasingabundanceofcheapnaturalgas.Thus,theonlybarrierstomeetingglobalenergyneedsbyPVtechnologyareeconomic;i.e.,improvementsineciencyandreductioninmanufacturingcostswillbeneededtoreachgrid-parity(whentheper-wattcostofPVenergyequalsthatofcoalandnaturalgas).Whilethinfilmsolarcellsstilldonotmeettheecienciesofsingle-crystalsiliconcells,theyarelessexpensivetomanufacture,andthefutureviabilityofsolarenergydependsonadvancesinthin-filmtechnology.Still,manyadvancesarerequiredtoreachgrid-parity;amajorgoalisthereductionofsolarcellcosttobelow1dollarperwatt.Themajorityofthinfilmsolarcellsusingcadmiumtellurideorcopper…indium…gallium…selenide(CIGS)areproducedusingchemicalvapourdeposition(CVD)orothersemiconductormanufacturingmethods;thesemethodsareenergy-intensiveandrequirespecializedequipment.Newmethodswillberequiredtoproduceinexpensiveflexiblethinfilmsolarcells.Oneoftheproblemsassociatedwithanycompoundsemi-conductorsolarcellisthatinordertohavesucientphotoncollectionitisnecessarytohavearelativelythickabsorberlayer(Fig.1a).However,athickabsorberlayerwillmeanthatthereisalargediusionlengthtothebackcontact,whichresultsinahigherpercentageofholeelectronrecombination,andhencelowereciency.Toovercomeboththeseissuesitwouldbedesirabletodesignanabsorberlayerthathasthebackcontactembeddedwithinthelayer(Fig.1b).Thiswouldallowforathickabsorberlayer,butarelativelyshortdiusionlength.Thiswill,intheory,allowforenhancedexcitondissociationandcarriertransport.Singlewalledcarbonnanotubes(SWNTs)haveahostofdesirablephysicalandelectronicpropertiesthatmakethempotentialasconductionchannelsandbackcontactmaterials.Inparticular,metallicSWNTsarecapableofballisticelectrontransportonmicronlengthscales,meaningtheyconductelectricalcurrentwithzeroenergylossfromscattering.Inaddition,theirDepartmentofChemistry,RiceUniversity,Houston,TX77005,USA.E-mail:arb@rice.edu;Tel:17133485610NatcoreTechnology,Inc.,87MapleAvenue,RedBank,NJ07701,USADepartmentofMechanicalEngineeringandMaterialsScience,RiceUniversity,Houston,TX77005,USACollegeofEngineering,SwanseaUniversity,SingletonPark,SwanseaSA28PP,Wales,UK.E-mail:A.R.Barron@swansea.ac.ukElectronicsupplementaryinformation(ESI)available:Photographicimagesofsamples,PVtestingconfiguration,andfullfabricationscheme.SeeDOI:10.1039/c3cp50435b‚Presentaddress:MaterialsScience&TechnologyDivision,LosAlamosNationalLaboratory,MSE549,POBox1663,LosAlamos,NM87545,USA.Received21stNovember2012,Accepted31stJanuary2013DOI:10.1039/c3cp50435bPC